I Built a Training Program That Beat Its Own Numbers by 15 Points

I built a training program that beat its own numbers by 15 points, then handed off the machine that did it.

Late last year I stood up an AI Skills Accelerator inside a nine-business-group organization. Fully optional. No chargeback for skipping it. The kind of program that usually limps to a 25 to 30 percent completion rate and everyone shrugs and calls that normal.

I didn't build a curriculum. I built the operating system underneath one. A dedicated sign-up site. Templatized communications so nothing got rebuilt from scratch every cohort. Automated workflows that cut my weekly prep time from 8 hours down to 3. A community space that kept people engaged between sessions instead of disappearing after the kickoff.

The program hit 40 percent completion. Enrollment crossed 400 participants.

And it didn't stay small. It scaled from one engineering team to three separate organizations in under a year, one of them over 60,000 people.

This is the pattern that's run through every program I've built. I find where the operational drag is, I build the infrastructure that removes it, and I leave it running better than I found it, whether or not I'm still in the room.

What's the bottleneck in your org that's quietly capping a good program's numbers?